Multiple Roles for the 3 in Vitro Reconstitution of Yeast

Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

In vitro reconstitution of yeast splicing with U4 snRNA reveals multiple roles for the 3 ′ stem−loop

Amy J. Hayduk, Martha R. Stark and Stephen D. Rader

RNA 2012 18: 1075-1090 originally published online March 12, 2012 Access the most recent version at doi:10.1261/rna.031757.111

References This article cites 38 articles, 20 of which can be accessed free at: http://rnajournal.cshlp.org/content/18/5/1075.full.html#ref-list-1 Email alerting Receive free email alerts when new articles cite this article - sign up in the box at the service top right corner of the article or click here

To subscribe to RNA go to: http://rnajournal.cshlp.org/subscriptions

In vitro reconstitution of yeast splicing with U4 snRNA reveals multiple roles for the 39 stem–loop

AMY J. HAYDUK,1 MARTHA R. STARK,1 and STEPHEN D. RADER2 Department of Chemistry, University of Northern British Columbia, Prince George, British Columbia, V2N 4Z9 Canada

ABSTRACT U4 small nuclear RNA (snRNA) plays a fundamental role in the process of premessenger RNA splicing, yet many questions remain regarding the location, interactions, and roles of its functional domains. To address some of these questions, we developed the first in vitro reconstitution system for yeast U4 small nuclear ribonucleoproteins (snRNPs). We used this system to examine the functional domains of U4 by measuring reconstitution of splicing, U4/U6 base-pairing, and triple-snRNP formation. In contrast to previous work in human extracts and Xenopus oocytes, we found that the 39 stem–loop of U4 is necessary for efficient base-pairing with U6. In particular, the loop is sensitive to changes in both length and sequence. Intriguingly, a number of mutations that we tested resulted in more stable interactions with U6 than wild-type U4. Nevertheless, each of these mutants was impaired in its ability to support splicing, indicating that these regions of U4 have functions subsequent to base pair formation with U6. Our data suggest that one such function is likely to be in tri-snRNP formation, when U5 joins the U4/U6 di-snRNP. We have identified two regions, the upper stem of the 39 stem–loop and the central domain, that promote tri-snRNP formation. In addition, the loop of the 39 stem–loop promotes di-snRNP formation, while the central domain and the 39-terminal domain appear to antagonize di-snRNP formation. Keywords: pre-mRNA splicing; in vitro reconstitution; U4 snRNA; di-snRNP formation; spliceosome activation; protein–RNA interaction; splicing complementation

INTRODUCTION and U6 share extensive sequence complementarity (Brow and Guthrie 1988; Guthrie and Patterson 1988), and their Splicing, an essential step in eukaryotic gene expression, interaction in the form of a di-snRNP is required for in- involves the removal of noncoding introns from precursor corporation of U6 into the active spliceosome (Brow and messenger RNA (pre-mRNA) and ligation of the flank- Guthrie 1988). U4, however, is not directly involved in splicing protein-coding exons to produce mature mRNA. This ing catalysis and is thought to dissociate from the spliceo- process is catalyzed by a large ribonucleoprotein complex, some prior to the first chemical reaction (Lamond et al. the spliceosome, composed of five small nuclear RNAs 1988; Yean and Lin 1991). Furthermore, there has been no (snRNAs)—U1, U2, U4, U5, and U6—and over 100 pro- direct biochemical identification of the proteins associated teins, some of which associate with a specific snRNA to form with U4 snRNA in the free snRNP, the sequences with the corresponding small nuclear ribonucleoprotein (snRNP). which they interact, or the roles they may perform during Through a highly ordered and tightly regulated series of in- snRNP and spliceosome assembly. teractions, these snRNPs assemble onto the pre-mRNA and, A powerful tool for examining the function and in- following a number of conformational changes leading to teractions of the U4 snRNP is in vitro reconstitution. This formation of the active spliceosome, catalyze the two chem- involves depleting endogenous U4 snRNA from cell extract ical steps of splicing. to abolish splicing activity, followed by complementation While each of the five snRNAs plays an essential role in with a modified version of U4 to examine the effects of splicing, that of U4 remains the least well understood. U4 mutations or investigate its interaction partners. Because U4 function is essential, in vitro reconstitution enables analysis of lethal mutations that would not be possible using in vivo 1These authors contributed equally to this work. methods. In vitro reconstitution assays have been developed 2 Corresponding author. for all five mammalian snRNPs (Wersig et al. 1992; Wolff and E-mail [email protected]. Article published online ahead of print. Article and publication date are Bindereif 1992; Se´gault et al. 1995; Will et al. 1996), as well as at http://www.rnajournal.org/cgi/doi/10.1261/rna.031757.111. yeast (Saccharomyces cerevisiae)U2,U5,andU6(Fabrizioetal.

RNA (2012), 18:1075–1090. Published by Cold Spring Harbor Laboratory Press. Copyright Ó 2012 RNA Society. 1075 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

Hayduk et al.

1989; McPheeters et al. 1989; O’Keefe et al. 1996); however, there has been no report of in vitro reconstitution of yeast U1 or U4, though the latter has been attempted (Horowitz and Abelson 1993). Reconstitution of mammalian U4 has provided valuable insight into the functionally important regions of this molecule. U4 snRNA can be divided into six domains: stem II, the 59 stem–loop, stem I, the central domain, the 39 stem–loop (the central stem–loop in mammalian U4), and the 39 terminal domain containing the Sm binding site (Fig. 1A). By examining deletion mutations within each of these domains, Wersig and Bindereif (1990, 1992) demonstrated that the 59 portion of U4, including stems I and II and the intervening 59 stem–loop, is required for U4/U6 di-snRNP formation, subsequent spliceosome assembly, and splicing in vitro, while the 39 portion of the molecule appeared to be dispensable. To investigate the domain structure and interactions of yeast U4, we have developed an assay for in vitro reconstitution of functional U4 snRNPs. By using DNA oligonucleotide-targeted degradation by RNase H, extract was specifically depleted of endogenous U4, abolishing splicing. Subsequent addition of in vitro transcribed (IVT) U4 snRNA restored up to 50% of the splicing ac- FIGURE 1. U4 degradation and consequent splicing inhibition require active splicing tivity of the extract. By using this sys- conditions. (A) Secondary structure of U4 (light gray) base-paired to U6 (dark gray) (Brow and Guthrie 1988) and as a free particle (Myslinski and Branlant 1991), with the targeting tem, we were able to define the mini- oligonucleotide binding site indicated (black line). (B) Extent of U4 degradation increases with mum functional sequence of yeast U4 actin concentration. Northern blot of U4 in splicing extract in the absence or presence of snRNA. Our results reveal that in con- targeting oligonucleotide and increasing amounts of unlabeled actin pre-mRNA. (C)U4 trast to mammalian U4, the 39 stem–loop degradation does not affect other snRNAs. Denaturing Northern blot of untreated splicing extract (SE, lane 1), mock-depleted extract (lane 2), and U4 depleted extract (DSE, lane 3), of yeast U4 is essential for interaction with probed for all five yeast snRNAs as indicated on the side. U5L indicates long U5; U5S, short U6 and thus for splicing. Strikingly, a U5; and U4-RNaseH, U4 cleavage product. (D) Extent of splicing inhibition increases with the number of U4 mutants base-paired to U6 concentration of cold actin in the U4 degradation step. Denaturing gel showing the generation more efficiently than wild-type U4. The of splicing intermediates and products after incubation of radioactively labeled actin pre- mRNA in mock-depleted (lane 3) or U4-depleted (lanes 4–8) extract preincubated with observation that these mutants were nev- increasing amounts of unlabeled actin pre-mRNA. Lanes 1 and 2 are standard splicing controls ertheless impaired in their ability to sup- incubated for 0 or 30 min, respectively. Lanes 3–8 are the same reactions as in panel B. Bands port splicing demonstrates that U4 plays corresponding to splicing substrate, intermediates, and products are indicated to the left of the an important role subsequent to U4/U6 gel (black and white boxes, 59 and 39 exons; black line, intron). The absolute splicing efficiency of each reaction is indicated below the gel. di-snRNP formation.

RESULTS U4 and thereby block the splicing activity of the extract. An effective technique for depletion of specific RNA molecules from extract is oligonucleotide-targeted degradation by The 59 end of U4 snRNA is efficiently degraded under RNase H. A previous attempt to develop a yeast U4 in conditions of active splicing vitro reconstitution assay used a complementary DNA To develop an in vitro assay for reconstitution of functional oligonucleotide to direct RNase H degradation of U4 yeast U4 snRNPs, we needed to inactivate the endogenous nucleotides 72–92 but was unsuccessful (Horowitz and

1076 RNA, Vol. 18, No. 5 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

A late function for the U4 39 stem–loop

Abelson 1993). This oligonucleotide had been shown to mock-depleted extract in the presence of 80 fmol pre-mRNA direct nearly complete degradation of the target sequence decreased the subsequent splicing activity of the extract by (Xu et al. 1990); however, degradation of this central region z20% compared with mock-depleted extract incubated in of U4 failed to abolish splicing activity (Horowitz and the absence of pre-mRNA (cf. lanes 2 and 3). Mock-depleted Abelson 1993), presumably because a functional portion of extract was generally able to splice z40%–70% of the labeled U4 remained intact. Here, we instead targeted the 59 end of actin pre-mRNA, while depletion of U4 consistently reduced U4 (Fig. 1A), as it has been shown to be essential for in- this number to <1%. teraction with U6 and, consequently, for splicing (Vankan et al. 1990; Wersig and Bindereif 1990, 1992). IVT U4 RNA restores splicing in U4-depleted extract Normally, the 59 end of U4 is found base-paired to U6, forming intermolecular stem II (Fig. 1A), and, as a result, Having identified conditions that allowed efficient depletion is almost completely resistant to oligonucleotide-targeted of endogenous full-length U4 and concomitant inhibition of RNase H degradation (Xu et al. 1990). To circumvent the splicing, we attempted to restore splicing activity by addition protection provided by U4/U6 stem II and to accomplish of IVT U4. Titration of IVT U4, from 28–2830 nM, showed efficient U4 degradation, we took advantage of the changes effective restoration of splicing relative to depleted extract in the U4 secondary structure that take place during the (Fig. 2A, cf. lanes 3–9 to lane 2). Maximum reconstitution splicing cycle. For the active spliceosome to be formed, base- was achieved by addition of 283 nM IVT U4 (lane 6). This pairing interactions between U4 and U6 must be broken, U4 concentration corresponds to about 300-fold more than resulting in the release of U4 from the spliceosome in the form of a free particle. In this U4 species, the 59 region normally engaged in base-pairing with U6 is believed to adopt a largely single-stranded conformation that should be more accessible to the targeting DNA oligonucleotide (Fig. 1A). To produce conditions of active splicing, unlabeled actin pre-mRNA was added to yeast splicing extract in the presence of ATP. Titration of actin pre-mRNA showed that increasing amounts of pre-mRNA allowed higher U4 degradation efficiencies, with 80 fmol pre-mRNA resulting in nearly complete depletion of full-length U4 (Fig. 1B). The amount of actin pre-mRNA necessary for efficient depletion of U4 was found to be extract-dependent, ranging from 15–80 fmol. Under these conditions, U4 snRNA is specifically depleted, and there is no effect on the other snRNAs (Fig. 1C). The remaining U4 fragment (nucleotides ½nt 31–160, labeled U4-RNaseH) is visible but is substantially degraded.

Degradation of the 59 end of U4 snRNA inhibits splicing in vitro A critical requirement for the development of a useful U4 in vitro reconstitution assay is that degradation of endogenous U4 must inhibit the splicing activity of the extract. To test the effectiveness of splicing inhibition, U4-depleted extract was incubated with radioactively labeled FIGURE 2. Reconstitution of splicing activity in U4-depleted extract pre-mRNA, followed by gel electrophoresis to separate by IVT U4 is not limited by the extent of base-paired U4/U6 the splicing substrate from any intermediates or products formation. (A) Maximum splicing reconstitution is obtained with formed. Near-complete depletion of the full-length U4 re- addition of 283 nM IVT U4. Denaturing splicing gel as in Figure 1D. sulted in complete inhibition of splicing in the extract rela- (Lane 1) Mock-depleted extract. (Lanes 2–9) Depleted extract with IVT U4 added as indicated above lanes. Absolute splicing efficiency of tive to the mock-depleted control (Fig. 1D, cf. lanes 3 and 8). each reaction is indicated below gel. (B) Extent of base-pairing with Given that the number of splicing cycles that can be per- U6 snRNA increases with added IVT U4. Non-denaturing solution formed by a specific extract is limited (Raghunathan and hybridization gel probed using a radioactively labeled oligonucleotide Guthrie 1998a), it was important to ensure that the splicing complementary to U6, showing the base-pairing status of U6. (Lane 1) Untreated splicing extract. (Lane 2) Mock-depleted extract. (Lanes 3–10) performed during the degradation of U4 was not compro- U4-depleted splicing extract with IVT U4 added as indicated. The mising the splicing potential of the extract. Incubation of percentage of U6 in base-paired complex is shown below the gel.

www.rnajournal.org 1077 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

Hayduk et al. endogenous U4 levels (see Discussion) (Brenner and Guthrie 2006). Addition of >283 nM IVT U4 resulted in decreased splicing; therefore, subsequent experiments used 283 nM U4. A range of degradation oligonucleotide concentrations, from 0.05–50 mM, was examined to identify the optimal concentration for depletion of full-length U4 while maintaining the ability of the depleted extract to restore splicing upon addition of IVT U4. This analysis showed that maximum reconstitution was achieved after depletion with 0.8 mM oligonucleotide (data not shown). To ascertain whether U4/U6 di- snRNP levels were limiting for splicing reconstitution, we measured the corresponding U4/U6 levels for each of the splicing reactions in Figure 2A (Fig. 2B). In contrast to the splicing efficiency, which peaked at 283 nM U4, U4/U6 levels continued to increase with increasing U4 to a maximum of 91% at 2830 nM (lane 10). In untreated splicing extract, 60% of U6 was base-paired (lane 1) compared with 67% in mock- depleted extract (lane 2). By compari- son,only16%ofU6wasbase-paired when we reconstituted with 283 nM U4 FIGURE 3. The 39 stem–loop of U4 is required for efficient base-pairing to U6. (A) Secondary (lane 7), suggesting that there was in- structure of U4 snRNA with the 39 ends of the truncation mutants indicated. (B)U4 sufficient di-snRNP to support maxi- truncations lacking the 39 stem–loop do not support reconstitution of splicing. Denaturing gel analysis of splicing in mock-depleted (lane 1) or U4-depleted (lanes 2–6) extract reconstituted mal splicing. Nevertheless, when 566 using wild-type (WT) or mutant IVT U4. Splicing substrate, intermediates, and products are nM U4 was added, yielding 58% base- indicated to the left of the gel, as in Figure 1. (C) The 39 stem–loop of U4 is required for base- pairing of U6 (lane 8), the splicing pairing to U6 at 283 nM U4. Non-denaturing solution hybridization gel analysis of the base- efficiency decreased (Fig. 2A, lane 7). pairing status of U6 in mock-depleted (lane 1) or U4-depleted (lanes 2–6) extract reconstituted using WT or mutant IVT U4. (D) U4 truncations can be driven into complexes with U6 by Further increases in U4 concentration increasing their concentration above 283 nM. Non-denaturing solution hybridization analysis resulted in further decreased splicing of extent of U4/U6 formation as in panel C. (Lane 1) 283 nM full-length IVT U4. Increasing efficiency. concentrations of truncated U4 mutants: 283, 566, 1132, and 2830 nM (1–142 only to 1132 nM).

Splicing activities of U4 39 truncation mutants To examine the effect of truncating U4 at nt 68, U4 1–68 transcript was added to extract depleted of endogenous U4, To examine the functional domains of yeast U4, we first de- and its ability to restore splicing was assessed. Unexpect- termined the minimum sequence that supports U4/U6 formation and splicing using our reconstitution assay with U4 edly, the addition of U4 1–68 to the U4-depleted extract mutants. Given that stems I and II and the 59 stem–loop are did not result in reconstitution of splicing activity (Fig. 3B, all required for efficient spliceosome assembly and splicing in lane 4), in contrast to the ability of wild-type IVT U4 to human extract and Xenopus oocytes (Vankan et al. 1990; restore splicing to 21% from 1% in the U4-depleted extract Wersig and Bindereif 1990, 1992), we tested a mutant con- (Fig. 3B, lanes 2 and 3; splicing and U4/U6 formation taining these three structural features, which consisted of nt efficiencies reported in the text are averages from at least 1–68. This mutant, U4 1–68, lacks the Sm binding site, the 39 three measurements; see Table 1). stem–loop, and most of the central domain (Fig. 3A); how- Lack of splicing reconstitution by U4 1–68 may be due to ever, it does contain the 59 portion of the central domain the absence of the 39 portion of the central domain (nt 69– identified by Wersig and Bindereif (1992) and Vankan et al. 90). While not essential for splicing, deletion of this sequence (1992) as being important for splicing activity. from mammalian U4 has an appreciable effect on splicing

1078 RNA, Vol. 18, No. 5 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

A late function for the U4 39 stem–loop

to function in splicing is not due to a lack TABLE 1. Di-snRNP formation and splicing reconstitution efficiencies of the U4 mutants of stability; rather, these mutants lack a d Di-snRNP KD Stability functionally essential component of U4. Mutant formationa Splicingb Splicingc (nM) Error R2 (%)

9 6 SE 30 81 8 Interaction of the U4 39 truncation Mock 54 6 9 DSE 0.6 6 0.6 mutants with U6 6 6 Wild type 100 100 21 8 800 47 0.88 45 18 Despite the fact that U4 truncation mu- 1–68 7 6 12 0 0 1518 264 0.59 26 6 14 1–90 0 6 1 0 0 1472 131 0.78 31 6 9 tants 1–68 and 1–90 contain the only 1–142 109 6 27 49 6 10 8 6 2 670 34 0.90 29 6 9 regions of U4 known to interact with D39SL 3 6 6 0 0 2192 175 0.82 68 6 21 U6, their inability to reconstitute splic- DUSL 35 6 17 6 6 216 1 1137 75 0.92 72 6 9 ing in U4-depleted extract could be D 6 6 6 6 138 105 57 10 221 650 68 0.69 44 7 caused by reduced or absent base-pair- D131–133 60 6 33 28 6 756 1 836 59 0.89 32 6 12 Dbulge 60 6 19 11 6 226 1 762 33 0.95 73 6 13 ing with U6 snRNA. To investigate this D102–130 24 6 22 4 6 116 1 1105 113 0.85 85 6 3 possibility, we measured the base-pairing D107–126 8 6 9 0 0 1666 157 0.86 68 6 2 status of U6. As expected, U4 deple- D66–90 227 6 108 0 0 390 34 0.88 47 6 16 tion produced extract in which U6 was D 6 6 6 6 69–90 168 48 6 310 470 53 0.70 51 10 found exclusively in the free snRNP Dloop(AAAAAA) 12 6 9 0 0 1287 107 0.94 62 6 21 Dloop(GGCUU) 15 6 4 0 0 1851 23 0.999 49 6 23 (Fig. 3C, lane 2). Both wild-type U4 Dloop(UUUUC) 11 6 2 0 0 1317 123 0.93 57 6 22 and U4 1–142 were able to reconstitute

a formation of the di-snRNP in U4-de- Di-snRNP formation values are based on the concentration of U4 that gave the highest z reconstitution of splicing using wild-type U4, 283 nM, and are given as a percent of wild- pleted extract to 20% of normal levels type di-snRNP formation. (Fig. 3C, lanes 3 and 6); however, in the b Splicing efficiencies are presented as a percentage of the wild type, measured separately in majority of experiments we found that each experiment, and are the average of at least three measurements 6 SD. cAbsolute splicing efficiencies. Values shown are the average over all experiments (n $ 3). U4 1–142 yielded more U4/U6 than the dStability was measured as a percentage of transcript remaining after incubation in depleted wild type (Table 1). U4 1–68 and 1–90 extract under splicing conditions. were unable to reconstitute di-snRNP formation at the concentration tested (Fig. 3C, lanes 4 and 5). efficiency (Wersig and Bindereif 1992). We therefore con- The reduced ability of U4 truncations to form U4/U6 structed a longer 39 truncation mutant containing the entire may be due to a kinetic defect in base pair formation or, central domain (U4 1–90) (Fig. 3A) and examined its ability alternatively, to reduced stability of the base-paired prod- to reconstitute U4-depleted extract. Surprisingly, this mu- uct. To test the first possibility, we allowed the reconstitu- tant was also unable to reconstitute splicing (Fig. 3B, lane 5). tion reactions to proceed for varying lengths of time up to The final U4 39 truncation mutant we examined, U4 1– 36 min. We found that the fraction of U4/U6 did not change 142, contains the 39 stem–loop in addition to the complete after 12 min, indicating that the reactions had reached central domain, stems I and II, and the 59 stem–loop and equilibrium (data not shown). To test the second possibility, thus lacks only the 39 terminal domain of U4, which con- we performed titrations of the truncated U4 mutants, from tains the Sm protein binding site (Fig. 3A). Consistent with 141 to 2830 nM, and subsequently examined U4/U6 levels. reconstitution studies in human nuclear extract (Wersig and When we fit a binding curve to the titration data to calculate Bindereif 1992), we found that the 39 terminal domain of an apparent KD (see Materials and Methods), we found a yeast U4 was not essential for splicing in vitro (Fig. 3B, lane significant difference between the mutants, ranging from 6), although the average level of splicing restored, deter- 1528 nM for 1–68 to 670 nM for 1–142, compared with mined from multiple measurements, was only 49% of that 800 nM for the wild type (Table 1). These results are con- restored by wild-type U4 (Table 1). sistent with the truncations affecting di-snRNP stability. The lack of splicing reconstitution by U4 1–68 and 1–90 Surprisingly, the 1–142 mutant appears to have a more stable could be due to the removal of a functionally essential interaction with U6 than wild-type U4, suggesting that the region of the molecule or to decreased stability of these Sm ring of U4 may antagonize base-pairing with U6. mutants in splicing extract. To investigate this second pos- The ability of U4 1–68 and 1–90 to base pair to U6 at sibility, we examined the stability of wild-type and mutant high concentration suggested that they may be able to recon- IVT U4 snRNA after incubation in U4-depleted splicing stitute splicing at these concentrations. We therefore tested extract. The three mutants were all less stable than wild-type splicing reconstitution using these mutants at concentrations IVT U4, with U4 1–142, which was active in the reconsti- from 283–2830 nM, which corresponded to U4/U6 levels tution assay, having an intermediate stability (Table 1).This from 0%–98% for 1–68 and 0%–72% for 1–90 (Fig. 3D, suggests that the inability of the shorter 39 truncation mutants lanes 2–9). Neither mutant was able to reconstitute splicing

www.rnajournal.org 1079 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

Hayduk et al. at any concentration tested (data not shown). U4 1–142 was similarly tested in a splicing reconstitution titration assay and, like the wild type, showed maximum splicing at 283 nM, despite increasing di-snRNP formation up to the highest concentration tested (1132 nM, yielding 93% di-snRNP; lane 12). We conclude that the inability of U4 1–68 and 1–90 to reconstitute splicing does not stem exclusively from their reduced ability to interact with U6, as they do not support splicing even under conditions where high levels of U4/U6 are formed. Intriguingly, these experiments point to the 39 stem–loop of U4 as an important determinant of U4/U6 di-snRNP formation.

Mutational analysis of the 39 stem–loop of U4 snRNA To directly test whether the 39 stem– loop is required for di-snRNP forma- FIGURE 4. U4 39 stem–loop mutations that form comparable levels of U4/U6 to the wild type tion, we constructed a mutant, D39SL, (WT) are nevertheless defective in splicing reconstitution. (A) Secondary structures of thatwasmissingtheentire39 stem– constructs used in this experiment: WT U4 39 stem–loop, deletion of the 11-nt bulge (Dbulge), the upper stem and bulge (DUSL), the bulge and lower stem (D102–130 ½WT loop), loop (nt 91–141) (see Fig. 3A) but still or the lower stem and loop (D107–126). Locations of deleted (138; 131–133) or mutated (loop contained the Sm binding site. Consis- mutations) nucleotides are shown in bold on the WT stem–loop. (B) Deletions or mutations at tent with the above experiments, this many positions in U4’s 39 stem–loop affect U4/U6 formation. Non-denaturing gel showing the mutant was unable to participate in di- base-pairing status of U6 in mock-depleted (lane 1) or U4-depleted (lanes 2–13) extract reconstituted using WT or mutant IVT U4. The positions of free U6 and the U4/U6 di-snRNA snRNP formation at the normal con- are indicated to the left of the gel. (C) Deletions and mutations in U4’s 39 stem–loop reduce centration (Fig. 4B, lane 4) and was also splicing reconstitution. Denaturing gel showing splicing activity in mock-depleted (lane 1)or unable to restore splicing at this con- U4-depleted (lanes 2–13) extract reconstituted using WT or mutant IVT U4 as in B. Splicing centration or any other up to 2830 nM substrate, intermediates, and products are indicated to the left of the gel, as in Figure 1. (Fig. 4C, lane 4; data not shown). In- app deed, D39SL had the highest KD of any U4 mutant tested, tion and splicing had been deleted. To identify these se- 2192 nM (Table 1). quences, we constructed three mutants containing smaller Given the unexpected importance of the 39 SL for di- deletions within the upper portion of the 39 stem–loop. Dele- snRNP formation and splicing, we undertook a more tion of U138, a highly conserved bulged nucleotide (Myslinski detailed mutational analysis to identify functionally impor- and Branlant 1991), resulted in better-than-wild-type inter- app tant regions within this domain. We constructed 10 U4 action with U6 (KD = 650 nM) but restored only 10% of mutants containing deletions or base mutations through- the splicing activity seen with wild-type U4 (lane 6; Table 1). out the 39 stem–loop (Fig. 4A) and analyzed the ability of In contrast, deletion of nucleotides 131–133, a mutation each to restore splicing and di-snRNP formation in U4- previously shown to produce a cold-sensitive phenotype in depleted extract. vivo (Hu et al. 1995b), resulted in levels of di-snRNP for- To investigate whether the nucleotides important for di- mation slightly below that of the wild type (836 nM) and snRNP formation are located in the upper portion of the 39 splicing levels that were 28% that of the wild type (lane 7; stem–loop, we examined the effects of a number of dele- Table 1). Finally, similar to D138, deletion of the 9-nt central tions in this region. We constructed a mutant, DUSL, in bulge of the 39 stem–loop, Dbulge (Fig. 4A), resulted in which the entire upper stem and central bulge were deleted, above-wild-type interaction with U6 (762 nM) but substan- leaving only the distal portion of the stem and loop (Fig. tially below-wild-type restoration of splicing (11% of the wild app 4A). While this mutant had a KD of 1137 nM for di- type) (lane 8; Table 1). snRNP formation, the splicing activity restored was <10% While deletions within the upper portion of the 39 stem– that of the wild type (Fig. 4B and C, lane 5; Table 1), in- loop identified nucleotides that contribute to splicing, none dicating that sequences important for both di-snRNP forma- of these mutations resulted in the strong block to di-snRNP

1080 RNA, Vol. 18, No. 5 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

A late function for the U4 39 stem–loop formation that was produced by deletion of the entire 39 stem–loop. To investigate whether a sequence essential for di-snRNP formation resides within the lower portion of the 39 stem–loop, we generated five mutants containing deletions within this region. Deletion of the central bulge and distal portion of the stem, leaving the two sides of the upper stem connected by the wild-type loop sequence, produced D102–130 (Fig. 4A). This mutant had an apparent KD for di- snRNP formation of 1105 nM and low, but above background, splicing (Fig. 4B,C lane 9; Table 1). Deletion of the entire lower stem and loop, D107–126 (Fig. 4A), severely app impaired di-snRNP formation (KD , 1666 nM) and splicing (lane 10; Table 1), indicating that a sequence essential for di-snRNP formation had been deleted. The observation that the apparent KD for U4/U6 formation of D102–130 was substantially lower than D107–126 suggested that the loop nucleotides of the 39 stem–loop are essential for di-snRNP formation. To test this hypothesis, we generated mutants in which the 6 nt of the loop were replaced. First, we asked whether the sequence of the loop was important by replacing the wild-type 39-UUUUCG-59 sequence with all As. This mutant formed no U4/U6 di- snRNP at the standard concentration and, consequently, did not support splicing (Fig. 4B,C, lane 12), indicating that nucleotide identity in the loop is important for U4/U6 formation. An alternative loop sequence from Trypanosoma brucei had previously been shown to form U4/U6 in vivo in the context of a chimeric U4 (Bordonne´ et al. 1990). We tested this loop sequence, 39-GGCUU-59, and found that, as with the all-adenosine loop, it supported neither base pair FIGURE 5. The central domain of U4 functions during U4/U6 formation nor splicing (lane 11). Finally, to test whether the formation and splicing. (A) Secondary structure of U4 (gray) with length of the loop was important, we made a version lacking targeting oligonucleotide binding site indicated (black line). (B)U4 requires its central domain to re-associate with U6. Non-denaturing G114 at the 59 end (39-UUUUC-59). Similarly to the other solution hybridization gel showing the base-pairing status of U6 in loop mutants, this version of U4 was inactive in base pair mock-depleted extract (lane 1) or extract depleted of U4 in the formation and splicing (lane 13). Although all of the 39 absence (lane 2) or presence (lane 3) of added unlabeled actin pre- stem–loop mutants were able to base pair with U6 at mRNA. The positions of base-paired wild-type (WT) U4/U6, U6 base- paired to the truncated 59 portion of U4 (U4*/U6), and free U6 are concentrations >283 nM, those that were unable to form di- indicated to the left of the gel. The fraction of U6 base-paired to U4 in snRNP or splice at 283 nM were also unable to splice at each reaction is indicated below the gel. (C) Degraded U4 supports higher concentrations (i.e., D39SL, D107–126, and all loop splicing when associated with U6. Denaturing gel showing splicing mutants; data not shown). Based on these experiments, the activity in mock-depleted extract (lane 1) or extract depleted of U4 in the absence (lane 2) or presence (lane 3) of pre-mRNA added during loop appears to be the most important portion of the 39 the U4 degradation step. Splicing substrate, intermediates, and stem–loop for both U4/U6 formation and splicing. products are indicated to the left of the gel, as in Figure 1. Absolute splicing efficiency of each reaction is indicated below the gel. (D) Deletion of U4’s central domain increases steady-state levels of U4/U6 Function of the central domain di-snRNP. U4/U6 analysis as in B of mock-depleted (lane 1)or depleted (lanes 2–5) extracts reconstituted with no U4 (lane 2)orU4 Having demonstrated that the 39 stem–loop of U4 was variants as shown. Note that U4/U6 migrates faster with the mutant required for productive U4/U6 formation (i.e., di-snRNP U4s due to their smaller size. (E) U4’s central domain is necessary for splicing. Splicing analysis of the reactions in panel D. formation that supports splicing), we sought to determine whether it might also have a function later in the splicing reaction. We attempted to remove the 39 portion of U4 tions (in the absence of added pre-mRNA as done previously) after di-snRNP formation using an oligonucleotide target- (Horowitz and Abelson 1993). ing the central domain (Fig. 5A). This resulted in separate In the absence of targeting oligonucleotide (Fig. 5B, lane but stable 59 and 39 portions of U4 (data not shown). De- 1), the majority of U6 is associated with U4. Similarly, fol- gradation was performed under either splicing conditions lowing cleavage of U4 under nonsplicing conditions (lane 2), (in the presence of added pre-mRNA) or nonsplicing condi- the majority of U6 remained associated with the 59 portion

www.rnajournal.org 1081 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

Hayduk et al. of U4 (containing nt 1–64), demonstrating that the complex of U6 with truncated U4 was stable. Cleavage of U4 under splicing conditions, however, resulted in the conversion of almost all U6 to the free snRNA (lane 3). We infer that the truncated U4 and U6 were separated during splicing and were unable to re-anneal. Surprisingly, extract containing the cleaved U4/U6 complex supported subsequent splicing of radioactively labeled pre-mRNA at z40% of normal splicing levels (Fig. 5C, lanes 1,2), while extract containing truncated U4 separated from U6 was not able to support splicing (lane 3). These results demonstrate that the 59 portion of U4 was able to remain in a stable, functional complex with U6 FIGURE 6. U4 deletion mutants are defective in triple-snRNP after central domain degradation and that the 39 stem–loop formation. Non-denaturing Northern blot analysis of unextracted does not need to be physically attached to the 59 end of U4 (i.e., protein-containing) samples. Untreated extract (lane 1), mock- (nt 1–64 in this experiment) after U4/U6 formation. Due to depleted extract (lane 2), and depleted extract reconstituted with nothing (lane 3) or with IVT U4 variants (283 or 566 nM, lanes 4– the stability of the 39 end of U4, these experiments were 17). Location of snRNP species indicated by labels at left.U4*/U6 unable to determine the step at which splicing is blocked in indicates aberrantly migrating U4/U6 di-snRNP seen in all recon- the absence of the 39 stem–loop. stituted extracts. These results raise the possibility that U4’s central domain (Figs. 1A, 3A) is not necessary for U4 function beyond its role in joining the two functionally essential ends of the (i.e., with proteins present) (Cheng and Abelson 1987; molecule. To test this, we reconstituted splicing with two U4 Raghunathan and Guthrie 1998a). Under these conditions, mutants, D66–90 and D69–90, in which the central domain all of the base-paired U6 is actually in tri-snRNP, and there was deleted. Wersig and Bindereif (1992) previously showed is no U4/U6 di-snRNP present (Fig. 6, lane 1; Raghunathan that the nucleotides in human U4 corresponding to yeast and Guthrie 1998a). Upon addition of ATP, tri-snRNP dis- positions 65–68 contributed to splicing in vitro, so we made sociates into free U4 and U6 snRNPs, which subsequently one deletion that removed these nucleotides along with the reform U4/U6 and tri-snRNP (lane 2). Depletion of U4 rest of the central domain (D66–90), and another (D69–90) from extract leads to the complete absence of U4/U6 and that retained them. Surprisingly, both mutants base-paired a strongly reduced signal at the position where tri-snRNP with U6 far more than wild-type U4 (Fig. 5D, lanes 3–5). runs (lane 3). The residual signal at the tri-snRNP position app Titrations yielded KD values of 390 and 470 nM, respec- may reflect incomplete degradation of U4 or may simply tively, compared with 800 nM for the wild type (Table 1), be nonspecific background; but in any case, there is no de- the most stable interactions with U6 of any U4 mutants that tectable splicing under these conditions (see Fig. 1D). we have tested. When the ability of these mutants to support All of the U4 mutants tested yielded less tri-snRNP after splicing was measured, however, D66–90 had no activity and reconstitution than did wild-type U4, which reconstituted D69–90 had very little activity, albeit consistently above 24% of the level of tri-snRNP seen in mock-depleted ex- depleted extract (Fig. 5E; Table 1), consistent with the tract. In particular, the five U4 mutants that had lower ap- observations of Wersig and Bindereif (1992). In contrast to parent KDs than the wild type, namely, 1–142, D66–90, the oligo degradation experiments (Fig. 5B,C), these deletion Dbulge, D138, and D69–90 (lanes 6–15), had 16%–27% as experiments suggest that the central domain contributes to much tri-snRNP as the wild type yet more than D39SL, splicing but antagonizes U4/U6 formation (see Discussion). which was comparable to the depleted lane (cf. lane 3 to lanes 16,17). The amount of tri-snRNP formed did not increase with additional U4 above 283 nM, in contrast to snRNP analysis of U4 tri-snRNP formation di-snRNP formation (e.g., cf. lanes 4 and 5), which is con- Certain U4 mutants described above were notable for having sistent with the maximum splicing activity occurring at 283 a lower apparent KD than wild-type U4 (U4 1–142, D138, nM, and may suggest that splicing is limited by tri-snRNP Dbulge, D66–90, and D69–90), yet none of them supported availability. splicing as well as the wild type. One possibility is that these This demonstrates that all of these mutants are defective U4 variants were defective in the next assembly step after in tri-snRNP formation. Notably, they support a wide range U4/U6 di-snRNP formation, namely, U4/U6.U5 tri-snRNP of splicing efficiencies, from 49% of the wild type for U4 formation. To test this, we measured the ability of these mu- 1–142% to 0% for D66–90. The lack of correlation between tants to assemble into tri-snRNPs at several concentrations tri-snRNP levels and splicing activity indicates that there (Fig. 6). is yet another step after tri-snRNP formation in which the SnRNP levels in extract can be analyzed on a non- 39 stem–loop and central domain of U4 play a role (see denaturing gel without phenol extracting the samples Discussion).

1082 RNA, Vol. 18, No. 5 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

A late function for the U4 39 stem–loop

An important observation is that the di-snRNP present DISCUSSION in the reconstituted lanes of the gel in Figure 6 (lanes 4–17) has a lower mobility than di-snRNP from mock-depleted Conditions for successful U4 reconstitution extract (lane 2). We have confirmed that this band, labeled U4*/U6 in the figure, does not contain U5 snRNA (data In this study, we have developed the first in vitro splicing not shown), so the aberrant mobility must be either due to reconstitution system for U4 snRNP in yeast. A key require- a difference in the protein complement compared to un- ment of this assay was to deplete endogenous U4 snRNA treated extract or due to a conformational change in one or under conditions of active splicing so that the 59 end of U4 more components (most likely U4). This may in turn be targeted for RNase H degradation was no longer base-paired responsible for the reduced levels of tri-snRNP and splicing to U6 and was therefore accessible to the targeting DNA activity upon reconstitution. oligonucleotide. We were able to reconstitute the splicing activity of the depleted extract by addition of IVT U4, which allowed us to discover an unexpected role for the Oligonucleotide accessibility of the 39 stem–loop 39 stem–loop in di-snRNP formation, as well as in later steps of U4 snRNA of spliceosome assembly (steps at which U4 39 domains may Mutation of specific regions of the 39 stem–loop of U4 may function are summarized in Fig. 8). inhibit di-snRNP formation or subsequent splicing by The amount of splicing reconstitution we observed was destabilizing an interaction with a protein. To investigate somewhat lower than that seen with other snRNAs (our whether the 39 stem–loop participates in protein/RNA in- range was 21%–76% of the mock reaction, compared with teractions, we examined the ability of oligonucleotides 50%–100% of untreated extract for U6 ½Fabrizio et al. 1989, complementary to the 59 or 39 side of this domain (Fig. 7A) and 55%–70% of the mock reaction for U2 ½McPheeters to direct degradation of U4 in extract by RNase H in the et al. 1989), and required higher concentrations of IVT U4 presence or absence of proteins. Surprisingly, the two sides (283 nM compared to 40–125 nM for U6 and U2 ½Fabrizio of this domain exhibited unequal oligonucleotide acces- et al. 1989; McPheeters et al. 1989; McGrail et al. 2006 and sibilities. Only the 59 side was efficiently degraded in the 125 nM for U5 ½McGrail and O’Keefe 2008). Approximately presence of proteins (Fig. 7B, lanes 2,3), and neither side 50% of the added IVT U4 was degraded in extract over the was susceptible to degradation in the absence of proteins course of splicing reconstitution. In addition, the aberrant (Fig. 7B, lanes 6,7). Because the Sm site of U4 is known mobility of the reconstituted di-snRNP on a non-denaturing to bind the Sm proteins, this region was used as a degrada- gel (Fig. 6) suggests that much of the IVT U4 is either mis- tion control. As expected, it was protected from degra- folded or does not associate properly with proteins. Never- dation in the presence of proteins but was accessible to theless, its ability to reconstitute tri-snRNP with normal the oligonucleotide, and therefore susceptible to degra- mobility, as well as splicing activity, suggests that some frac- dation, in the absence proteins (Fig. 7B, lanes 4,8). These tion of the U4 interacts normally with the rest of the splicing results suggest that the 39 stem–loop of U4 interacts with machinery. at least one protein. The decrease in reconstitution efficiency seen with IVT U4 concentrations >283 nM may be due to increased com- petition for U4-specific proteins, effectively diluting the proteins available such that fewer U4 RNA molecules receive the full protein complement required for functional- ity. Alternatively, an inhibitory component in the IVT solution (such as residual acrylamide) could be responsible for the inhibition. Conditions for complete depletion of endogenous U4 from extracts had to be determined for each extract tested, requiring 15–80 fmol of actin, compared with 4 fmol for a standard splicing reaction. This may reflect the need to cycle the entire pool of endogenous U4 through the splicing pathway, perhaps multiple times, to ensure complete degra- FIGURE 7. Proteins confer asymmetric accessibility to U4’s 39 stem– dation. We cannot explain the variability between extracts. It loop in untreated splicing extract. (A) Secondary structure of the U4 39 stem–loop showing the binding sites of oligonucleotides targeting was also surprising that U4 depletion was only successful the 59 and 39 sides (black and gray lines, respectively). (B) The U4 39 under conditions of active splicing, as ATP alone is known stem–loop is only degraded in the presence of proteins. Denaturing to cause the dissociation of tri-snRNP in yeast as well as Northern blot probed for U4, showing U4 size in the absence (lanes human extract (Berget and Robberson 1986; Raghunathan 1,5) or presence (lanes 2–4,6–8) of oligonucleotide targeting the 59 or 39 side of the 39 stem–loop, or the Sm site. Lanes 1–4 are in the and Guthrie 1998b). This suggests the possibility that U4 presence of proteins, while lanes 5–8 are in the absence of proteins. released from the spliceosome is in a different conformation

www.rnajournal.org 1083 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

Hayduk et al.

does not significantly inhibit di-snRNP formation (Vankan et al. 1990; 1992; Wersig and Bindereif 1990; 1992). In- deed, these studies further demonstrated this domain to be dispensable for spliceosome assembly and even for splicing, although efficiency was reduced compared with wild-type U4. Similarly, a series of point mutations throughout the 39 stem–loop of yeast U4 were found to be functional in vivo, although deletion of nt 131–133 did result in a cold- sensitive phenotype (Hu et al. 1995b). In contrast to these studies suggesting that the 39 stem–loop is not a functionally important element of U4 snRNA, an in vivo analysis by Bordonne´ et al. (1990) found that substitution of the yeast U4 39 stem–loop with the comparable region of T. brucei U4, an 11-nt stem–loop, decreased interaction with U6 by 60%–70% and was lethal. Simi- larly, the 39 stem–loop of human U4atac snRNA has been shown to be required for in vivo activity (Shukla et al. 2002). Hence, the requirement for the U4 39 stem–loop appears to vary between or- ganisms and even between the major and minor spliceosome. Furthermore, while the 39 stem–loop of yeast U4 is clearly necessary for di-snRNP formation, there appears to be considerable flexibility in the sequence requirements FIGURE 8. Model of functional domains in the 39 portion of U4 and the spliceosome of this domain. assembly steps at which they function. The region of U4 that promotes di-snRNP formation is The U4 39 stem–loop mutants can be indicated in black, the region that increases tri-snRNP formation is in dark gray, the region that decreases di-snRNP levels is boxed in white, and the region that both decreases di-snRNP divided into two categories based on and increases tri-snRNP is boxed in light gray. In the lower panel, the portions of U4 tested in their di-snRNP formation and splicing this work that we propose to be functionally important are shown in bold, with arrows reconstitution efficiencies: those that indicating the points in the splicing cycle at which they act. restored reduced levels of di-snRNP and splicing, and those that restored in- from U4 released from tri-snRNP under nonsplicing condi- creased levels of di-snRNP but reduced levels of splicing. tions, with the 59 end of the former being more accessible We note that the two classes of 39 stem–loop mutants appear than the latter. Another possibility is that the difference in to segregate by location in the stem–loop: Those that in- oligo accessibility is due to differences in protein composi- crease the apparent KD are in the lower stem–loop, while tion of the two types of free U4 snRNP. those that have no significant effect (D131–133) or reduce the apparent KD (D138 and Dbulge) are in the upper stem. Two lines of evidence demonstrate the importance of the The 39 stem–loop of yeast U4 is required for efficient loop nucleotides of U4’s 39 stem–loop (nt 114–119) in di- di-snRNP formation snRNP formation. First, deletion of either the upper stem Reconstitution by U4 39 truncation mutants showed that (DUSL) or the lower stem (D102–130) has only a modest the 39 stem–loop plays a critical role in promoting the in- effect on the apparent KD for U4/U6 formation. Impor- teraction with U6 that is required for subsequent recon- tantly, neither of these changes the loop sequence. In stitution of splicing. This was unexpected, as a number of contrast, D107–126, which removes the loop along with previous mutational analyses of U4 in human nuclear extract the lower stem, dramatically increases the apparent KD. and Xenopus oocytes have shown that deletion of this domain Second, mutations in the loop itself have a greater impact

1084 RNA, Vol. 18, No. 5 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

A late function for the U4 39 stem–loop on U4/U6 formation than deletion of either the upper or for the observed elevation of di-snRNP levels is that these lower stem, and one of the loop mutants has the second- mutations strengthen interactions with a splicing factor, such highest apparent KD of all of the mutations tested, as Prp24, that promotes di-snRNP formation. Such gain- exceeded only by deletion of the entire 39 stem–loop. of-function mutations are, however, relatively rare, and it Giventhatdi-snRNPmustformpriortospliceosome would be surprising to find even one, let alone two, in the assembly, any mutations that strongly reduce di-snRNP relatively limited survey of mutations described here. Alter- formation will also block splicing. Consistent with this, natively, these mutations may disrupt an interaction that the only mutations tested that eliminated splicing were would normally lead to destabilization of U4/U6, namely, those that changed the 39 loop, with the exception of unwinding by the Brr2 helicase prior to spliceosome ac- D66–90, which removes central domain nucleotides pre- tivation. In this case, we would expect to see an accumu- viously shown to be important for splicing (Wersig and lation of spliceosomes or multi-snRNP particles (larger than Bindereif 1992). di-snRNP) (Raghunathan and Guthrie 1998b), yet we do The loop mutations we have tested indicate that both the not. The remaining possibility is that these residues may length and the sequence of the loop are important. Like- play a role in formation or stabilization of the U4/U6.U5 wise, the chimeric U4 with an altered loop sequence dis- tri-snRNP. We favor this role for the nucleotides on the 39 cussed above had a strong effect on di-snRNP formation in side of the upper stem–loop because our analysis of snRNP vivo (Bordonne´ et al. 1990). The even more dramatic re- particles shows an accumulation of di-snRNP and reduced duction in U4/U6 levels seen when we placed this loop tri-snRNP. One mechanism by which this region of the 39 sequence (Dloop(GGCUU)) on the U4 wild-type stem may stem–loop could promote tri-snRNP formation could be reflect differences between the in vitro and in vivo assays through interaction with a U5 protein. used or may be a consequence of differences in the stem. In contrast, one of the U4 point mutations analyzed in vivo by The 39 side of the 39 stem–loop may bind a protein Hu et al. (1995b), G114U, falls within the loop of the 39 stem–loop. Alone, this mutation did not produce a phe- Our oligonucleotide accessibility data showed strong pro- notype, suggesting either that our deletion of this nu- tein dependence. For an RNA sequence to be degraded by cleotide (Dloop(UUUUC)) impacts function more than RNase H, it must be available for base-pairing with a mutations at this position or that the in vitro assay is complementary DNA oligonucleotide. Binding of this tar- more sensitive. geting oligonucleotide may be blocked if the RNA sequence is involved in secondary or tertiary structure interactions or in protein binding. In the absence of proteins, the 39 The 39 stem–loop functions at a late stage stem–loop of yeast U4 is expected to adopt a highly stable of spliceosome assembly secondary structure with an estimated melting temperature Previous studies of U4 function have mostly focused on its of 77°C (Owczarzy et al. 2008). Given that the estimated role in di-snRNP formation; however, there is evidence melting temperatures of the oligonucleotides targeting the indicating that it plays a role in subsequent steps, such as tri- 59 and 39 sides of the 39 stem–loop are substantially lower snRNP formation. In yeast, humans, and Xenopus, deletion (45°Cand43°C, respectively), it is not surprising that they of the 59 stem–loop results in wild-type or better di-snRNP were not able to direct RNase H degradation in the absence formation but invariably inhibits tri-snRNP formation and of proteins. splicing (Bordonne´ et al. 1990; Wersig and Bindereif 1992; The differential accessibility of the two sides of the 39 Wersig et al. 1992). We also tested the effect of deletion stem–loop in the presence of proteins was unexpected. The of U4’s 59 stem–loop in our reconstitution system and accessibility of the 59 side demonstrates that the binding of similarly found increased levels of U4/U6, but no splicing proteins to U4 changes the secondary structure of the 39 (data not shown). The 59 stem–loop of U4 is associated stem–loop. This may result from a protein binding directly with Prp4, which has been implicated in tri-snRNP forma- to the 39 side of the stem, disrupting base-pairing within tion (Bordonne´ et al. 1990). In addition, mutations in the the stem and thereby protecting the 39 side while releas- second WD repeat of Prp4 block dissociation of U4 from ing the 59 side for base-pairing with a targeting oligonu- the spliceosome, suggesting yet another role for U4 just cleotide. Notably, a recent crystal structure of the U4 core prior to spliceosome activation (Ayadi et al. 1997). (the Sm ring with U4’s Sm binding site and flanking stem– An unexpected late function (i.e., after di-snRNP for- loops) supports the possibility of a protein binding the mation) of U4’s 39 stem–loop is suggested by the second protected side of the 39 stem–loop (Leung et al. 2011). For class of 39 stem–loop mutants. These mutations, D138 and example, the structure shows that the conserved bulged U Dbulge, restored similar levels of di-snRNP formation as in the 39 stem–loop (U114 in humans, U138 in S. cerevisiae) the wild-type sequence and reduced apparent KDs, yet re- is positioned away from the Sm ring such that it would be duced splicing, indicating that they impair a process sub- accessible to a protein. The asymmetric pattern of oligonu- sequent to base-pairing with U6. One possible explanation cleotide accessibility is also consistent with the asymmetry of

www.rnajournal.org 1085 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

Hayduk et al. mutational effects observed in previous studies (Hu et al. data in Figure 5C demonstrate that separating the 59 and 39 1995b). portions of U4 reduces splicing by z60%, suggesting that the remaining 40% of a standard in vitro splicing yield is catalyzed prior to recycling. Importantly, cleavage of the Role of U4’s central domain central domain of U4 thus provides a practical method to As discussed above, our data indicate an unexpected role study single turnover splicing kinetics. for the 39 stem–loop after di-snRNP formation. Surpris- ingly, the 39 stem–loop does not have to be connected to Role of the 39 terminal domain the 59 portion of U4 to carry out this role. Extract in which the central domain of U4 has been completely degraded, Our titration analysis of U4 allowed us to make quantita- leaving stable 59 and 39 portions of the snRNA, is still able tive comparisons between the mutants. For example, it is to support splicing (Fig. 5C; Horowitz and Abelson 1993), notable that deletion of the 39 stem–loop gave a higher implying that once the di-snRNP has formed, separation of apparent KD than truncation of U4 at nt 68 (2192 nM vs. the two halves of U4 does not prevent any additional in- 1518 nM), which appears to suggest that removing all of U4 teractions that may occur between them. Surprisingly, dele- after nt 68 has less impact on U4/U6 formation than simply tion of the central domain (nt 66–90 or 69–90) results in deleting the 39 stem–loop. The observation that U4 1–90 higher steady-state levels of U4/U6, suggesting that the has a similar apparent KD to U4 1–68 indicates that the central domain antagonizes U4/U6 formation. The higher lower KD of U4 1–68, compared with that of D39SL, is not levels could be due to faster di-snRNP formation, slower due to the absence of the central domain and must there- tri-snRNP formation, or slower unwinding during spliceo- fore be due to the absence of the 39 terminal domain (nt some activation (Fig. 8). Given the accumulation of U4/U6 143–160). We conclude that the 39 terminal domain serves on snRNP gels and the relatively low levels of tri-snRNP to reduce steady-state levels of U4/U6, because its presence after reconstitution with central domain deletion mutants, in D39SL increases the apparent KD relative to U4 1–90. Our we favor the interpretation that the most important func- observation that U4 1–142 has a lower apparent KD than tion of the central domain occurs during addition of U5 to wild-type U4 is consistent with this conclusion. the U4/U6 di-snRNP. Taken together, these experiments The function of the 39 terminal domain in decreasing demonstrate that during U4/U6 formation, the central U4/U6 levels could be due to direct antagonism of the rate domain is only required to keep the 39 stem–loop physically of di-snRNP formation or to acceleration of U4/U6 un- attached to U4, while during tri-snRNP formation, the winding during spliceosome assembly. Our snRNP gel anal- central domain is required to maintain appropriate spacing ysis, which shows no change in tri-snRNP levels between any between the 59 region and the 39 stem–loop of U4. of the mutants, is more consistent with the former possi- Our results are consistent with previous work on the bility, although it is unclear how the 39 terminal domain, central domain. In Xenopus, deletion of the central domain which includes the Sm protein binding site, would slow di- had no effect on U4/U6 formation but blocked splicing snRNP formation. (Vankan et al. 1990). Genetic experiments in yeast also The data presented here reveal the unexpected role suggested that the requirements for the central domain are played by the 39 stem–loop of yeast U4 in promoting U4/ flexible: Deletion of nt 62–88 had only a mild cold-sensitive U6 formation and tri-snRNP assembly, with the loop phenotype (Hu et al. 1995b), and insertion of an extra 76 nt nucleotides promoting di-snRNP formation and with the into the central domain had no effect on growth (Hu et al. upper stem and central domain promoting tri-snRNP for- 1995a). mation (Fig. 8). The 39 terminal domain, which includes More detailed analysis of the central domain has shown the Sm protein binding site, appears to slow the rate of di- that only the 59 end has a strong sequence requirement. snRNP formation. Interestingly, a recent article (He et al. In Xenopus, substitution of the first 3 nt (64–66) blocked 2011) reports that a mutation in the analogous stem of splicing, but no other 3-nt substitution in the central do- human U4atac is linked to microcephalic osteodysplastic main had a strong effect (Vankan et al. 1992). Similarly, primordial dwarfism type I (MOPD I), also known as small deletions throughout the central domain of human Taybi-Linder Syndrome (TALS). Our data suggest that the U4 revealed that only the 59 end (nt 64–67) was important U4atac mutation may cause problems in U4/U6 formation for efficient splicing, whereas other deletions reduced splic- or in subsequent tri-snRNP formation. These results em- ing by at most 50% (Wersig and Bindereif 1992). phasize the need for structural information to reveal the The ability of the two halves of U4 in the splicing extract nature of the conformational changes that occur during to separate, and the inability of the separated halves to re- RNA structural rearrangements, such as in the transition associate with U6, allows us to estimate the fraction of splic- from free U4 to the U4/U6 di-snRNP. A thorough un- ing in a normal assay that occurs in the initial round and the derstanding of this conformational change will require both fraction that occurs after recycling of U6 (i.e., dissociation static structures as well as information about the dynamic and reassembly) (Raghunathan and Guthrie 1998a,b). The processes that occur during this transition.

1086 RNA, Vol. 18, No. 5 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

A late function for the U4 39 stem–loop

MATERIALS AND METHODS column (Edge BioSystems) according to the manufacturer’s instructions. DNA was ethanol precipitated and quantified using a DNA oligonucleotides NanoDrop spectrophotometer (NanoDrop Technologies). Wild- type U4 and U4 mutants were transcribed in vitro from plasmid or U4-targeting oligonucleotides, primers used in site-directed mu- PCR template using the MEGAshortscript Kit (Ambion) accord- tagenesis, and U4 and U6 probes are listed in Table 2. ing to the manufacturer’s instructions. RNA was gel purified and quantified as described for PCR products. Synthesis of U4 snRNA and actin pre-mRNA Unlabeled actin pre-mRNA was synthesized from HindIII- linearized JPS149 plasmid (Vijayraghavan et al. 1986; Mayas et al. Wild-type U4 was synthesized from StyI-linearized pT7U4 plas- 2006) using the MEGAshortscript Kit (Ambion) according to the mid template (Ghetti et al. 1995). Templates for synthesis of the manufacturer’s instructions and was gel purified and quantified as U4 mutants were generated by site-directed mutagenesis from described above. Radioactively labeled actin pre-mRNA was pT7U4 plasmid. The same forward primer containing the T7 transcribed and purified as previously described (Aukema et al. promoter, oligonucleotide A, was used for all mutants, and the 2009). reverse primers are listed in Table 2. PCR products were purified by electrophoresis through a denaturing polyacrylamide gel and Oligonucleotide-directed RNase H degradation of U4 visualized by ultraviolet shadowing. Following homogenization in a 1.5-mL microcentrifuge tube using a small pestle (Kontes Whole-cell extract was prepared from protease-deficient yeast strain Scientific), the gel slice was mixed with 200 mLdH2O and incu- BJ2168 as described (Ansari and Schwer 1995). U4 degradation bated for 10 min at 70°C. The gel slurry was passed over a DTR reactions were performed in 8 mLcontaining60mMKPO4 (pH 7.0),

TABLE 2. DNA oligonucleotides used in this study

Sequence (59 to 39) Mutant

PCR primers A AATTAATACGACTCACTATAGGGATCCTTATGCACGGGAAATAa All B TTTCAACCAGCAAAAACACA 1–68 C GACGGTCTGGTTTATAATTAAATTTCA 1–90 D CCCTACATAGTCTTGAAGTATTCA 1–142 E AAAGGTATTCCAAAAATTGACGGTCTGGTTTATAATTAAATTTCA D39SL F AAAGGTATTCCAAAAATTAGTATTCAAAAGCGAACACCGACGGTCTGGTTTATAATTAAATTTCA DUSL G AAAGGTATTCCAAAAATTCCCTCATAGTCTTGAAGTATTC D138 H AAAGGTATTCCAAAAATTCCCTACATATTGAAGTATTCAAAAGCG D131–133 I AAAGGTATTCCAAAAATTCCCTACATAGTCAGTATTCAAAAGCGAACACCGACCATGAGGAGACGG Dbulge J AAAGGTATTCCAAAAATTCCCTACATAGTCAAAAGCGACCATGAGGAGACGG D102–130 K AAAGGTATTCCAAAAATTCCCTACATAGTCTTGAGAATTGACCATGAGGAGAC D107–126 L AAAGGTATTCCAAAAATTCCCTACATAGTCTTGAAGTATTCTTTTTTGAACACCGAATTGACCAT Dloop (AAAAAA) M AAAGGTATTCCAAAAATTCCCTACATAGTCTTGAAGTATTCAAGCCGAACACCGAATTGACCAT Dloop (GGCUU) N AAAGGTATTCCAAAAATTCCCTACATAGTCTTGAAGTATTCGAAAAGAACACCGAATTGACCAT Dloop (UUUUC) O59-AGATTGTGTTTTTGCTGGTTGTCCTCATGGTCAATTCGG (inside fwd) D66–90 59-CCGAATTGACCATGAGGACAACCAGCAAAAACACAATCT (inside rev) 59-AAAGGTATTCCAAAAATTCCCTAC (outside rev) P59-TGTGTTTTTGCTGGTTGAAATCCTCATGGTCAATTCGG (inside fwd) D69–90 59-CCGAATTGACCATGAGGATTTCAACCAGCAAAAACACA (inside rev) 59-AAAGGTATTCCAAAAATTCCCTAC (outside rev) U4-targeting oligonucleotides 59end CTGATATGCGTATTTCCCGTGCATAAGGAT Central domain CGGTCTGGTTTATAATTAAATTTC 59 side of 39 SL TCAAAAGCGAACACCGAATTGACC 39 side of 39 SL ATAGTCTTGAAGTATTCAAAAGCG Sm site AGGTATTCCAAAAATTCCCTAC Probes U1 CAGTAGGACTTCTTGATC U2 (L15) CAGATACTACACTTG 59 U4 (CM6) TCAACCAGCAAAAACACAATCTCG 39 U4 (14B) AGGTATTCCAAAAATTCCCTAC U5 (7SmWTNR) AAGTTCCAAAAAATATGGCAAGC U6–467 ATTGTTTCAAATTGACC U6–6D AAAACGAAATAAATCTCTTTG

aT7 promoter indicated in bold.

www.rnajournal.org 1087 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

Hayduk et al.

2.5 mM MgCl2, 3% PEG 8000, 50% (v/v) yeast extract, 10 mM phenol/chloroform, ethanol precipitated, and electrophoresed ATP, 0.8 mM targeting oligonucleotide (59 end or central domain) through a 6% (19:1) 7 M urea denaturing polyacrylamide gel (Table 2), and 15–80 fmol unlabeled IVT actin pre-mRNA (depend- for 1 h at 400 V. The gel was then exposed to a phosphor screen ing on the splicing extract). No exogenous RNase H was added as at 80°C. yeast extract contains sufficient endogenous RNase H activity. To calculate splicing efficiency, autoradiograms were visualized Reactions were incubated for 30 min at 30°C. Oligonucleotide ac- and quantified as described above. The percentage of splicing was cessibility of the 39 stem–loop and Sm site was examined as de- calculated by dividing the intensity of bands corresponding to scribed above except 15 mM targeting oligonucleotide (Table 2) was product (mRNA and lariat) by the total intensity of the starting used. For oligonucleotide accessibility of deproteinized U4, yeast material plus the product (pre-mRNA, lariat, and mRNA). The extract was phenol/chloroform extracted, ethanol precipitated, and level of splicing activity restored by the U4 mutants was expressed resuspended in 23 RNase H buffer (40 mM Tris-HCl at pH 7.5, as a percentage of that of wild-type U4 by subtracting the resi- 10 mM MgCl2, 4 mM DTT, 0.012% BSA). Four microliters of this dual splicing activity found in the U4-depleted reaction from resuspension, containing approximately the amount of U4 found in both the mutant and wild-type U4 activities and dividing the 4 mL extract, was mixed with targeting oligonucleotide, 2 U RNase former value by the latter. Splicing efficiencies were determined H (Ambion), and 10 mM ATP to a volume of 8 mL and incubated at least in triplicate. as described above. Analysis of U4/U6 base-pairing Northern blot analysis of U4 degradation and IVT U4 stability Following the U4 degradation reaction, IVT U4 (2.4 pmol for single reactions, or up to 24 pmol in titrations) was added, and the U4 degradation reactions were terminated by addition of 200 mL mixture was incubated for 12 min at 23°C to allow interaction stop buffer (0.3 M sodium acetate, 1 mM EDTA, 0.1% SDS, with U6 to occur. Two hundred microliters of stop buffer was 34 mg/mL Escherichia coli tRNA). Reactions were phenol/chloro- then added, and the samples were phenol/chloroform extracted form extracted, ethanol precipitated, and electrophoresed through a and ethanol precipitated. To examine the effect of degradation 6% (19:1) 7 M urea denaturing polyacrylamide gel in 13 TBE at of the U4 central domain on U4/U6 interaction, stop buffer was 400 V for 30 min. RNA was transferred to a nylon membrane added immediately following the degradation reaction. Pre- (Hybond+, GE Healthcare; Owl semi-dry electroblotter) in 13 cipitated RNA was resuspended in 10 mLhybridizationbuffer TBE for 20 min at 2.5 mA/cm2, which was then probed using (150 mM NaCl, 50 mM Tris-HCl at pH 7.4, 1 mM EDTA) and a 32P-labeled probe complementary to the 39 (Fig. 1B,C) or 59 incubated for 15 min at 23°C with z200–500 fmol (100,000 cpm) (Fig. 7B) portion of U4 (Table 2) and exposed to a phosphor 32P-labeled U6–467 probe (Table 2). Samples were electrophor- screen. esed through a precooled 9% (29:1) non-denaturing polyacryl- Stability of wild-type and mutant U4 transcripts was tested in amide gel in 13 TBE for 1 h at 4°C at 300 V. The gel was exposed the presence of depleted splicing extract. We added 2.4 pmol IVT to a phosphor screen at 80°C. U4 to a U4 degradation reaction after the 30°C incubation, and it To calculate base-pairing efficiency, autoradiograms were was incubated for 12 min at 23°C. Then 4 fmol cold actin tran- visualized and quantified as described above. Base-pairing effi- script was added and incubated for an additional 30 min at 23°C ciency was calculated by dividing the intensity of the U4/U6 band to simulate the conditions of a splicing reaction. The reactions by the total intensity of both bands (U4/U6 and free U6). The level were then treated as above, and the Northern blot was probed of base-pairing activity restored by the U4 mutants was expressed with the 59 U4 probe (Table 2). For control reactions, 200 mL stop as a percentage of that of wild-type U4 by subtracting the residual buffer was added to 2.4 pmol of each IVT U4 in the absence of base-pairing found in the U4-depleted reaction from both the depleted splicing extract and processed as above. Stability was cal- mutant and wild-type U4 activities and dividing the former value culated as the percentage of U4 transcript remaining relative to by the latter. Base-pairing efficiencies were determined at least in the control. triplicate. Autoradiograms were visualized using a Cyclone Phosphor- imager and OptiQuant software (Packard Instruments). Image Calculation of apparent KD values contrast and brightness were linearly adjusted using Adobe Photo- for U4/U6 formation shop to enhance clarity. Absolute values for di-snRNP formation were plotted as a function Pre-mRNA splicing assay of the concentration of added IVT U4 in Kaleidagraph (Synergy Software, version 4.1.2), and the data were fit to the curve U4/U6 2 2 2 2 To assess the effect of U4 degradation on splicing activity, 4 fmol (%) = (100 3 ½U4 )/(½U4 +KD ). Values for the error and R internally 32P-GTP-labeled actin pre-mRNA in vitro transcript were calculated in Kaleidagraph. As is frequently found in binding was added to the degradation reaction and incubated for 30 min experiments, apparent KD fits were apparently cooperative, and at 23°C. To assess reconstitution of splicing in U4-depleted extract allowing a free parameter for the Hill coefficient yielded a value by wild-type or mutant U4, 2.4 pmol IVT U4 (unless otherwise close to 2. We therefore set the Hill coefficient to exactly 2 for our indicated) was added following the U4 depletion reaction and fits to minimize the number of free parameters. In addition, the incubated for 12 min at 23°C to allow snRNP assembly to occur apparent KD clearly results from several competing processes, prior to addition of internally labeled actin pre-mRNA and a including at least the rate of U4/U6 formation and the rate of further incubation for 30 min at 23°C. Splicing reactions were Brr2-mediated U4/U6 unwinding. It should not be interpreted as terminated by addition of 200 mL stop buffer, extracted with a true equilibrium constant for U4/U6 formation. Nevertheless,

1088 RNA, Vol. 18, No. 5 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

A late function for the U4 39 stem–loop

the U4 titration data demonstrate that U4/U6 formation behaves snRNA, a component of the minor spliceosome, in the develop- as a simple equilibrium, which justifies using the apparent KD to mental disorder MOPD I. Science 332: 238–240. compare the ability of U4 variants to base pair with U6. Horowitz DS, Abelson J. 1993. A U5 small nuclear ribonucleoprotein particle protein involved only in the second step of pre- mRNA splicing in Saccharomyces cerevisiae. Mol Cell Biol 13: SnRNP analysis 2959–2970. Hu J, Field D, Wang A, Friesen JD. 1995a. The central region of the To assess the formation of tri-snRNP in depleted extracts re- yeast U4 snRNA is not important for hammerhead catalysis, but constituted with U4, 1 mM spermidine was added to the U4/U6 may act as a domain spacer. Nucleic Acids Symp Ser 1995: 25–28. base-pairing reconstitution reaction. Protein/RNA complexes were Hu J, Xu D, Schappert K, Xu Y, Friesen JD. 1995b. Mutational analysis of Saccharomyces cerevisiae U4 small nuclear RNA iden- electrophoresed through a precooled 4% (80:1) native polyacryl- tifies functionally important domains. Mol Cell Biol 15: 1274– amide gel in 13 TGM buffer (50 mM Tris base, 50 mM glycine, 1285. 2mMMgCl2) for 3 h at 4°C at 160 V (Raghunathan and Guthrie Lamond AI, Konarska MM, Grabowski PJ, Sharp PA. 1988. Spliceo- 1998a). Gels were transferred to nylon membrane in 20 mM NaPO4 some assembly involves the binding and release of U4 small (pH 6.4), 20 min at 2.5 mA/cm2, and then probed with 32P-labeled nuclear ribonucleoprotein. Proc Natl Acad Sci 85: 411–415. U6-6D (Table 2). Leung AKW, Nagai K, Li J. 2011. Structure of the spliceosomal U4 snRNP core domain and its implication for snRNP biogenesis. Nature 473: 536–539. Mayas RM, Maita H, Staley JP. 2006. Exon ligation is proofread by the ACKNOWLEDGMENTS DExD/H-box ATPase Prp22p. Nat Struct Mol Biol 13: 482–490. McGrail JC, O’Keefe RT. 2008. The U1, U2 and U5 snRNAs crosslink We thank K. Aukema for technical assistance and H. Oh for initial to the 59 exon during yeast pre-mRNA splicing. Nucleic Acids Res experiments with loop mutants, as well as other members of the 36: 814–825. Rader laboratory and Re´my Bordonne´ for helpful discussions McGrail JC, Tatum EM, O’Keefe RT. 2006. Mutation in the U2 and comments on the manuscript. This work was supported by snRNA influences exon interactions of U5 snRNA loop 1 during NSERC Discovery Grant 298521 to S.D.R., NSERC Canada Grad- pre-mRNA splicing. EMBO J 25: 3813–3822. McPheeters DS, Fabrizio P, Abelson J. 1989. In vitro reconstitution of uate Scholarship 347734, MSFHR Trainee Award 857(06-1)BM, functional yeast U2 snRNPs. Genes Dev 3: 2124–2136. two UNBC Health Research Project Awards, and a UNBC Research Myslinski E, Branlant C. 1991. A phylogenetic study of U4 snRNA Project Award to A.J.H. reveals the existence of an evolutionarily conserved secondary structure corresponding to ‘‘free’’ U4 snRNA. Biochimie 73: 17–28. Received December 7, 2011; accepted January 26, 2012. O’Keefe RT, Norman C, Newman AJ. 1996. The invariant U5 snRNA loop 1 sequence is dispensable for the first catalytic step of pre- mRNA splicing in yeast. Cell 86: 679–689. REFERENCES Owczarzy R, Tataurov AV, Wu Y, Manthey JA, McQuisten KA, Almabrazi HG, Pedersen KF, Lin Y, Garretson J, McEntaggart NO, Ansari A, Schwer B. 1995. SLU7 and a novel activity, SSF1, act during et al. 2008. IDT SciTools: a suite for analysis and design of nucleic the PRP16-dependent step of yeast pre-mRNA splicing. EMBO J acid oligomers. Nucleic Acids Res 36: W163–W169. 14: 4001–4009. Raghunathan PL, Guthrie C. 1998a. A spliceosomal recycling factor Aukema KG, Chohan KK, Plourde GL, Reimer KB, Rader SD. 2009. that reanneals U4 and U6 small nuclear ribonucleoprotein Small molecule inhibitors of yeast pre-mRNA splicing. ACS Chem particles. Science 279: 857–860. Biol 4: 759–768. Raghunathan PL, Guthrie C. 1998b. RNA unwinding in U4/U6 Ayadi L, Miller M, Banroques J. 1997. Mutations within the yeast U4/ snRNPs requires ATP hydrolysis and the DEIH-box splicing factor U6 snRNP protein Prp4 affect a late stage of spliceosome assembly. Brr2. Curr Biol 8: 847–855. RNA 3: 197–209. Se´gault V, Will CL, Sproat BS, Lu¨hrmann R. 1995. In vitro re- Berget SM, Robberson BL. 1986. U1, U2, and U4/U6 small nuclear constitution of mammalian U2 and U5 snRNPs active in splicing: ribonucleoproteins are required for in vitro splicing but not Sm proteins are functionally interchangeable and are essential for polyadenylation. Cell 46: 691–696. the formation of functional U2 and U5 snRNPs. EMBO J 14: Bordonne´ R, Banroques J, Abelson J, Guthrie C. 1990. Domains of 4010–4021. yeast U4 spliceosomal RNA required for PRP4 protein binding, Shukla GC, Cole AJ, Dietrich RC, Padgett RA. 2002. Domains of snRNP–snRNP interactions, and pre-mRNA splicing in vivo. human U4atac snRNA required for U12-dependent splicing in Genes Dev 4: 1185–1196. vivo. Nucleic Acids Res 30: 4650–4657. Brenner TJ, Guthrie C. 2006. Assembly of Snu114 into U5 snRNP Vankan P, McGuigan C, Mattaj IW. 1990. Domains of U4 and U6 requires Prp8 and a functional GTPase domain. RNA 12: 862– snRNAs required for snRNP assembly and splicing complemen- 871. tation in Xenopus oocytes. EMBO J 9: 3397–3404. Brow DA, Guthrie C. 1988. Spliceosomal RNA U6 is remarkably Vankan P, McGuigan C, Mattaj IW. 1992. Roles of U4 and U6 conserved from yeast to mammals. Nature 334: 213–218. snRNAs in the assembly of splicing complexes. EMBO J 11: 335– Cheng SC, Abelson J. 1987. Spliceosome assembly in yeast. Genes Dev 343. 1: 1014–1027. Vijayraghavan U, Parker R, Tamm J, Iimura Y, Rossi J, Abelson J, Fabrizio P, McPheeters DS, Abelson J. 1989. In vitro assembly of yeast Guthrie C. 1986. Mutations in conserved intron sequences affect U6 snRNP: a functional assay. Genes Dev 3: 2137–2150. multiple steps in the yeast splicing pathway, particularly assembly Ghetti A, Company M, Abelson J. 1995. Specificity of Prp24 binding of the spliceosome. EMBO J 5: 1683–1695. to RNA: a role for Prp24 in the dynamic interaction of U4 and U6 Wersig C, Bindereif A. 1990. Conserved domains of human U4 snRNAs. RNA 1: 132–145. snRNA required for snRNP and spliceosome assembly. Nucleic Guthrie C, Patterson B. 1988. Spliceosomal snRNAs. Annu Rev Genet Acids Res 18: 6223–6229. 22: 387–419. Wersig C, Bindereif A. 1992. Reconstitution of functional mammalian He H, Liyanarachchi S, Akagi K, Nagy R, Li J, Dietrich RC, Li W, U4 small nuclear ribonucleoprotein: Sm protein binding is not Sebastian N, Wen B, Xin B, et al. 2011. Mutations in U4atac essential for splicing in vitro. Mol Cell Biol 12: 1460–1468.

www.rnajournal.org 1089 Downloaded from rnajournal.cshlp.org on April 16, 2012 - Published by Cold Spring Harbor Laboratory Press

Hayduk et al.

Wersig C, Guddat U, Pieler T, Bindereif A. 1992. Assembly and Wolff T, Bindereif A. 1992. Reconstituted mammalian U4/U6 snRNP nuclear transport of the U4 and U4/U6 snRNPs. Exp Cell Res 199: complements splicing: a mutational analysis. EMBO J 11: 345–359. 373–377. Xu Y, Petersen-Bjørn S, Friesen JD. 1990. The PRP4 (RNA4) protein Will CL, Ru¨mpler S, Klein Gunnewiek J, van Venrooij WJ, Lu¨hrmann of Saccharomyces cerevisiae is associated with the 59 portion of the R. 1996. In vitro reconstitution of mammalian U1 snRNPs U4 small nuclear RNA. Mol Cell Biol 10: 1217–1225. active in splicing: the U1-C protein enhances the formation of Yean SL, Lin RJ. 1991. U4 small nuclear RNA dissociates from a yeast early (E) spliceosomal complexes. Nucleic Acids Res 24: 4614– spliceosome and does not participate in the subsequent splicing 4623. reaction. Mol Cell Biol 11: 5571–5577.

1090 RNA, Vol. 18, No. 5